hadoop-common-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hong Tang (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HADOOP-6226) Create a LimitedByteArrayOutputStream that does not expand its buffer on write
Date Wed, 02 Sep 2009 08:32:32 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-6226?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12750329#action_12750329

Hong Tang commented on HADOOP-6226:

@Jothi, thanks for taking time porting the class to io package and writing unit tests for
it. Some comments for consideration:

- The comment of the test case refers to the class LimitedByteArrayOutputStream.
+/** Unit tests for LimitedByteArrayOutputStream */

- Should put some random contents in the input byte array. Something like the following (capitalize
variable INPUT because it is now static final):
private static final byte[] INPUT = new byte[SIZE];
static {
    new Random().nextBytes(INPUT);

- Suggest to write only one byte to test whether an exception is thrown due to limit exceeded.

- Suggest to test that after reset() and writing INPUT, writing one more byte would cause
an exception.

- Suggest to do reset(SIZE-1) instead of reset(SIZE/2).

- Suggest to add an assertion (getLimit() == SIZE) after reset() and (getLimit() == SIZE-1)
after reset(SIZE-1).

> Create a LimitedByteArrayOutputStream that does not expand its buffer on write
> ------------------------------------------------------------------------------
>                 Key: HADOOP-6226
>                 URL: https://issues.apache.org/jira/browse/HADOOP-6226
>             Project: Hadoop Common
>          Issue Type: New Feature
>    Affects Versions: 0.20.0
>            Reporter: Jothi Padmanabhan
>            Assignee: Jothi Padmanabhan
>             Fix For: 0.21.0
>         Attachments: hadoop-6226-v1.patch, hadoop-6226-v2.patch, hadoop-6226.patch
> For MAPREDUCE-318, we would benefit from a ByteArrayOutputStream that would support a
fixed size buffer that will not expand dynamically on writes. 

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message